Coronaviral Main Protease Induces LPCAT3 Cleavage and Endoplasmic Reticulum (ER) Stress

Zoonotic coronaviruses infect mammals and birds, causing pulmonary and gastrointestinal infections. Some animal coronaviruses, such as the porcine epidemic diarrhea virus (PEDV) and transmissible gastroenteritis virus (TGEV), lead to severe diarrhea and animal deaths. Gastrointestinal symptoms were also found in COVID-19 and SARS patients. However, the pathogenesis of gastrointestinal symptoms in coronavirus diseases remains elusive. In this study, the main protease-induced LPCAT3 cleavage was monitored by exogenous gene expression and protease inhibitors, and the related regulation of gene expression was confirmed by qRT-PCR and gene knockdown. Interestingly, LPCAT3 plays an important role in lipid absorption in the intestines. The Mpro of coronaviruses causing diarrhea, such as PEDV and MERS-CoV, but not the Mpro of HCoV-OC43 and HCoV-HKU1, which could induce LPCAT3 cleavage. Mutagenesis analysis and inhibitor experiments indicated that LPCAT3 cleavage was independent of the catalytic activity of Mpro. Moreover, LPCAT3 cleavage in cells boosted CHOP and GRP78 expression, which were biomarkers of ER stress. Since LPCAT3 is critical for lipid absorption in the intestines and malabsorption may lead to diarrhea in coronavirus diseases, Mpro-induced LPCAT3 cleavage might trigger gastrointestinal symptoms during coronavirus infection.

In this study, we report that Mpro induces LPCAT3 cleavage and ER stress. Interestingly, this cleavage was only found in viruses that caused gastrointestinal symptoms. Considering the critical role of LPCAT3 in lipid absorption and macrophage polarization, it is suggested that Mpro-induced LPCAT3 cleavage might lead to gastrointestinal symptoms in coronavirus diseases, especially severe diarrhea. LPCAT3 deficiency might be a common feature for highly virulent coronavirus infection.

Recombinant Vectors and Transfection
Plasmid encoding SARS-CoV-2 (Wuhan strain) Mpro was stored in the lab. Full-length cDNA of MERS CoV Mpro, PEDV Mpro, HCoV-HKU1 Mpro, and HCoV-OC43 Mpro were synthesized (Sangon Biotech, Shanghai, China). LPCAT3 cDNA was purchased from Addgene. Target DNA sequences were amplified using high-fidelity 2 × PrimeStar max DNA polymerase (Takara, Dalian, China), and additional restriction enzyme sites as well as the additional tags. Amplified DNA fragments were digested by BamH I and XhoI, and then cloned into pcDNA3.1 by T4 ligase (NEB, Beijing, China). The SARS-CoV-2 Mpro (H41Y/C145S) mutant was generated using Mut Express II Fast Mutagenesis Kit V2 (Vazyme Biotech, Nanjing, China). For lentiviral vectors, full-length PEDV Mpro with a Flag-tag at the N-terminus was cloned into a pLVX-Puro vector. All constructions were validated by DNA sequencing, and the details of the plasmid sequence can be found in supplementary materials File S1.

Western Blot and qRT-PCR
Cells were lyzed using RIPA buffer (Cell Signaling Technology, Boston, USA) at indicated time points. Twenty micrograms of total cellular lysate were loaded in each lane for SDS-PAGE gel electrophoresis. Separated proteins were transferred to the PDVF or NC membrane for Western blot analysis (Cytiva, Shanghai, China). Antibodies for β-actin (#4967), FLAG (#8146S), and MYC (#2276S) were purchased from Cell Signaling Technology. LPCAT3 (#ab232958) antibody was purchased from Abcam (Cambridge, UK). GAPDH (KC-5G4) antibody was purchased from KANGCHEN LLC. (Chengdu, China).
To quantify the mRNA expression, total RNA was extracted using TRIzol reagent (Takara), and the first-strand cDNA was generated using HiScript III RT SuperMix for qPCR

Data Process
Experiments were performed at least in three independent biological replicates and results were represented as mean ± SEM. Student's t-test was used for statistical analyses.
In cellular lysate from SARS-CoV-2-infected HEK293T-hACE2 cells, an unexpected band (~37 kDa) was noticed when the LPCAT3 (~55 kDa) was analyzed, suggesting potential LPCAT3 cleavages during SARS-CoV-2 infection ( Figure 1A). A slight decrease was observed for full-length LPCAT3 ( Figure 1A). To test whether Mpro leads to the cleavage, a Flag-tag was fused to the C-terminus of SARS-CoV-2 Mpro in a pcDNA3.1 vector to generate pcDNA3.1-SCV2 Mpro ( Figure 1B), since additional tags in the N-terminus of Mpro would dramatically reduce its catalytic activity. Interestingly, the cleaved LPCAT3 was clearly found with the expression of exogenous SARS-CoV-2 Mpro in HEK293T cells, indicating that Mpro was enough to induce LPCAT3 cleavage ( Figure 1C).  Moreover, the LPCAT3 antibody used here was polyclonal. It was unclear whether the unexpected band was a truncated form of LPCAT3 or an off-target of the polyclonal antibody. To exclude the potential off-targets, a Myc-tag was fused to the N-terminus of LPCAT3 ( Figure 1B). Notably, the truncated Myc-tagged LPCAT3 was observed when SARS-CoV-2 Mpro was expressed in cells ( Figure 1D). This confirmed that exogenous SARS-CoV-2 Mpro could induce LPCAT3 cleavage. Afterward, increased LPCAT3 cleavage was found following the accumulation of exogenous SARS-CoV-2 Mpro in cells ( Figure 1E), and the cleavage was dose-dependent on the expression level of Mpro ( Figure 1F).

Mpro-Induced LPCAT3 Cleavage Might Be Related to the Gastrointestinal Symptoms in Coronavirus Diseases
SARS-CoV-2 infects various cell types causing different symptoms [5,34,35]. The crosstalk between viral proteases and host proteins is diverse in different cell types. To test whether SARS-CoV-2 Mpro would induce LPCAT3 cleavage in all types of cells, pcDNA3.1-SCV2 Mpro was used to transfect a series of cell lines, including human colorectal cancerderived HCT116 cells (Figure 2A), human cervical cancer-derived HeLa cells ( Figure 2B), human hepatocellular cancer-derived HuH-7 cells ( Figure 2C), and human breast cancerderived MCF-7 cells ( Figure 2D). LPCAT3 expression was detected in all tested cells ( Figure 2). As expected, cleaved LPCAT3 was found in all tested cells with exogenous SARS-CoV-2 Mpro expression, even though the expression level of Mpro was low in MCF-7 cells because of the poor transfection efficiency ( Figure 2). It demonstrated that the SARS-CoV-2 Mpro-induced LPCAT3 cleavage was independent of cell types.  Figure 2D). LPCAT3 expression was detected in all tested cells ( Figure 2). As expected, cleaved LPCAT3 was found in all tested cells with exogenous SARS-CoV-2 Mpro expression, even though the expression level of Mpro was low in MCF-7 cells because of the poor transfection efficiency (Figure 2). It demonstrated that the SARS-CoV-2 Mpro-induced LPCAT3 cleavage was independent of cell types. Most coronaviruses cause no or mild symptoms after infection exampling of HCoV-HKU1 and HCoV-OC43. However, some coronaviruses cause severe diseases, such as PEDV and MERS-CoV. LPCAT3 is abundant in the liver, intestines, and adipose tissue [26,28,36]. Since the important role of LPCAT3 in the intestine was mentioned earlier, it was hypothesized that Mpro-induced LPCAT3 cleavage might lead to gastrointestinal symptoms in coronavirus diseases. To test it, the Mpro of MERS-CoV, PEDV, HCoV-HKU1, and HCoV-OC43 was inserted into pcDNA3.1 vector with a Flag-tag at the C-terminus as SARS-CoV-2 Mpro, respectively. After verification by DNA sequencing, these recombinant plasmids were used to transfect HEK293T cells. Mpro of MERS-CoV and PEDV-induced LPCAT3 cleavage ( Figure 3A,B). Even though the expression level of HCoV-HKU1 and HCoV-OC43 Mpro was relatively high, no LPCAT3 cleavage was observed ( Figure 3C,D). It suggested that the Mpro-induced LPCAT3 might be related to the gastrointestinal symptoms of coronavirus diseases. Most coronaviruses cause no or mild symptoms after infection exampling of HCoV-HKU1 and HCoV-OC43. However, some coronaviruses cause severe diseases, such as PEDV and MERS-CoV. LPCAT3 is abundant in the liver, intestines, and adipose tissue [26,28,36]. Since the important role of LPCAT3 in the intestine was mentioned earlier, it was hypothesized that Mpro-induced LPCAT3 cleavage might lead to gastrointestinal symptoms in coronavirus diseases. To test it, the Mpro of MERS-CoV, PEDV, HCoV-HKU1, and HCoV-OC43 was inserted into pcDNA3.1 vector with a Flag-tag at the C-terminus as SARS-CoV-2 Mpro, respectively. After verification by DNA sequencing, these recombinant plasmids were used to transfect HEK293T cells. Mpro of MERS-CoV and PEDV-induced LPCAT3 cleavage ( Figure 3A,B). Even though the expression level of HCoV-HKU1 and HCoV-OC43 Mpro was relatively high, no LPCAT3 cleavage was observed ( Figure 3C,D). It suggested that the Mpro-induced LPCAT3 might be related to the gastrointestinal symptoms of coronavirus diseases.

Mpro Indirectly Induced LPCAT3 Cleavage
The Mpro protein is highly conserved and cleaves similar motifs. However, the Mpro of SARS-CoV-2, MERS-CoV, and PEDV could induce LPCAT3 cleavage (Figures 1 and 3), but the Mpro of HCoV-OC43 and HCoV-HKU1 could not (Figure 3). It hinted that the

Mpro Indirectly Induced LPCAT3 Cleavage
The Mpro protein is highly conserved and cleaves similar motifs. However, the Mpro of SARS-CoV-2, MERS-CoV, and PEDV could induce LPCAT3 cleavage (Figures 1 and 3), but the Mpro of HCoV-OC43 and HCoV-HKU1 could not (Figure 3). It hinted that the Mpro-induced cleavage of LPCAT3 might be indirect. GC376 was reported to covalently inhibit SARS-CoV-2 Mpro [37]. However, adding GC376 into the culture medium failed to abolish Mpro-induced LPCAT3 cleavage ( Figure 4A). Afterwards, the proteasome is the major protein degradation machinery in eukaryotic cells [38]. MG132 was a promised reversible proteasome inhibitor, which had been proven to inhibit SARS-CoV-2 Mpro in our early studies [39]. Unexpected, Mpro-induced LPCAT3 cleavage was still observed in the presence of MG132 (Figure 4B), suggesting a proteasome-independent cleavage. Moreover, the 41st amino acid residue histone and the 145th amino acid residue cysteine of Mpro formed a catalytic dyad to hydrolyze substrates. To confirm that the catalytic activity of Mpro was not required for LPCAT3 cleavage, a Mpro mutant was constructed, in which the His41 was replaced by tyrosine and the Cys145 was replaced by serine as previously described [40]. Not surprisingly, Mpro mutation still induced LPCAT3 cleavage as well as wild-type SARS-CoV-2 Mpro ( Figure 4C). Altogether, it clearly indicated that Mpro indirectly induced LPCAT3 cleavage.

Exogenous Mpro Caused ER Stress via LPCAT3 Cleavage
PEDV caused severe diarrhea in piglets, and notable gastrointestinal symptoms have been observed in 20-50% of patients infected with SARS-CoV-2 and MERS-CoV. It suggests that the gastrointestinal tract is an unignorable organ during coronaviral infection. LPCAT3 preferentially introduces polyunsaturated acyl onto the sn-2 position of lysophosphatidylcholine, modulating the membrane fluidity and playing an important role in lipoprotein production in the liver and intestine [26]. In addition, LPCAT3 knockdown

Exogenous Mpro Caused ER Stress via LPCAT3 Cleavage
PEDV caused severe diarrhea in piglets, and notable gastrointestinal symptoms have been observed in 20-50% of patients infected with SARS-CoV-2 and MERS-CoV. It suggests that the gastrointestinal tract is an unignorable organ during coronaviral infection. LPCAT3 preferentially introduces polyunsaturated acyl onto the sn-2 position of lysophosphatidylcholine, modulating the membrane fluidity and playing an important role in lipoprotein production in the liver and intestine [26]. In addition, LPCAT3 knockdown in the liver exacerbated ER stress and inflammation [28]. Then, would Mpro-induced LPCAT3 cleavage induce ER stress in the intestine cells? Exogenous SARS-CoV-2 Mpro was expressed in HCT116 cells, which would induce LPCAT3 cleavage (Figure 2A). The C/EBP homologous protein (CHOP) expression was obviously higher in cells transfected with pcDNA3.1-SCV2 Mpro compared to cells transfected with empty vectors ( Figure 5A). GRP78, another ER stress marker, was also increased in cells expressing exogenous SARS-CoV-2 Mpro ( Figure 5B). Interestingly, HCoV-OC43 Mpro failed to induce LPCAT3 cleavage ( Figure 3D), which did not boost the CHOP expression ( Figure 5C). An increased CHOP expression was found when the LPCAT3 expression was knocked down by siRNA ( Figure 5D,E). It demonstrated that Mpro-induced ER stress was related to LPCAT3 cleavage. Moreover, IPEC-J2 cells are porcine intestinal enterocytes, which are hard to transfect. A lentiviral vector encoding the PEDV Mpro was constructed to transduce IPEC-J2 cells ( Figure 5F). PEDV Mpro was successfully and stably expressed in IPEC-J2 cells ( Figure 5G) and increased CHOP mRNA was observed ( Figure 5H).

Discussion
Coronaviruses infecting livestock and humans are zoonotic viruses. There is evi-

Discussion
Coronaviruses infecting livestock and humans are zoonotic viruses. There is evidence indicating that HCoV-229E, -NL63, -OC43, and -HKU1 may originate from bats, rodents, cattle, and camels [41]. SARS-CoV and SARS-CoV-2 seem likely to have recently speciated from bat coronaviruses [42]. Considering the cross-species transmission history of coronaviruses, it is time to move the focus from human coronaviruses and take eyes on animal coronaviruses. Because of the phylogenetic conservation and the evolutionary history of coronaviruses, animal coronaviruses share many common features with human coronaviruses [43,44]. Many animal coronaviruses cause gastrointestinal symptoms. Although the respiratory tract symptoms caused by SARS-CoV-2 are a typical feature of COVID-19 [45], gastrointestinal symptoms have been observed in 20-50% of patients infected by SARS-CoV-2 [46]. Moreover, fatigue and gastrointestinal symptoms are typical features of long COVID [47], which may be related to the permanent damage in the gastrointestinal tracts by viral infection. Here, our studies indicated that Mpro indirectly cleaved LPCAT3, which was abundant in the intestine and essential for lipid absorption. Since malabsorption could lead to diarrhea, it suggested that the Mpro-induced LPCAT3 cleavage might play an important role in the gastrointestinal symptoms of coronavirus diseases.
LPCATs are essential to lipid metabolism and homeostasis. Based on their protein sequence, four LPCATs have been identified and classified into two families. LPCAT1 and LPCAT2 are members of the acylglycerophosphate acyltransferase family, which contains four conserved domains designated as LPA acyltransferase motifs. LPCAT3 and LPCAT4 belong to the membrane-bound O-acyltransferase (MBOAT) family, which contains MBOAT motifs but lacks the LPA acyltransferase motifs [26]. However, these LPCATs display distinct tissue distributions, enzymatic activities, and substrate preferences [48]. LPCAT3 is widely expressed and abundant in the testis, kidney, and metabolic tissues including the liver, intestine, and adipose [28,36]. In addition to the lysophosphatidylcholine (lysoPC) acyltransferase activity, LPCAT3 exhibits acyltransferase activities for lysophosphatidylethanolamine (lysoPE) and lysophosphatidylethanolamine (lysoPS) as substrates [49]. Most importantly, each LPCAT exhibits different acyl-CoA preferences. LPCAT3 prefers polyunsaturated fatty acyl CoAs (18:2-acyl-CoA or 20:4-acyl-CoA) as substrates [48]. Thus, LPCAT3 deficiency in the intestine reduces polyunsaturated phospholipid content and membrane fluidity, impairs passive fatty acid transport across the apical membrane of enterocytes, and decreases chylomicron assembly and secretion [27]. In this case, Mpro-induced LPCAT3 cleavage might lead to or aggravate gastrointestinal symptoms during viral infection. Since SARS-CoV-2, MERS-CoV, and PEDV Mpro could induce LPCAT3 cleavage but HCoV-HKU1 and HCoV-OC43 could not, it is reasonable to see diarrhea in patients and animals infected with SARS-CoV-2, MERS-CoV, and PEDV.
Recent studies have demonstrated that SARS-CoV-2 Mpro cleaves multiple host proteins, such as NF-κB essential modulator (NEMO) [21], Nod-like receptor family pyrin domain-containing protein 12 (NLRP12), transforming growth factor-β activated kinase 1 binding protein 1 (TAB1) [50], and signal transducer and activator of transcription 1 (STAT1) [19]. In this study, we have newly discovered SARS-CoV-2 Mpro-induced LPCAT3 cleavage. Unfortunately, the cleavage site and the mechanism are still elusive. There was a potent cleavage site in the N-terminus of human LPCAT3, which might be cleaved by Mpro. However, LPCAT3 might not be cleaved at that site since the cleaved fragment was too small (Figure 1). Moreover, LPCAT3 cleavage was independent of the catalytic activities of Mpro ( Figure 4B). Scientists have put a lot of effort into the interaction be-tween SARS-CoV-2 proteins and host proteins, which might be genetically associated with comorbidities of severe illness and long COVID [14]. Compared to other viral proteins, fewer host proteins have been identified to physically interact with proteases. Notably, we were convinced that SARS-CoV-2 Mpro and PLpro cleaved many host proteins, such as STAT1, TAB1, NLRP12, and NEMO. However, we failed to find them in these interactome studies. It is possible that the viral protease interacts and cleaves host proteins transiently, so it is difficult to identify associated proteins using physical interaction-based approaches, such as immunoprecipitation, pull-down assay, or yeast two-hybrid assay. Due to the elusive understanding of LPCAT3 cleavage during SARS-CoV-2 infection, it is still not clear what the biological impact of LPCAT3 cleavage is. Whether there are other host proteins involved in the cleavage of LPCAT3 by Mpro needs further investigation. Our preliminary results indicated that exogenous Mpro interrupted the lipid metabolism process in cells. However, the polyunsaturated phospholipid content and membrane fluidity of cells should be monitored in the future.
Importantly, diarrhea is the typical symptom of many animal coronavirus infections and leads to death. However, the antiviral agent remdesivir failed to alleviate diarrhea during PEDV and TGEV infection, even though these compounds efficiently inhibited viral replication. Our preprinting work indicated that a Mpro inhibitor stopped diarrhea in PEDV-infected piglets in 48 h, but viral genomic RNA was still detected until 7 days post-medicine management. This suggests that Mpro might directly cause diarrhea in coronavirus diseases, but the mechanism was still largely unclear. Our study suggested that Mpro-induced LPCAT3 cleavage might be critical to the gastrointestinal symptoms in coronavirus diseases, providing a hypothesis to explain how Mpro inhibitor could stop diarrhea before the clearance of viruses. Based on our studies, should inhibitors against LPCAT3 cleavage be screened for animal coronaviruses?